Identifying the genetic basis of variation in complex traits is a grand challenge. Genome-wide association studies suggest that many alleles with small effects may be responsible for many common diseases. When allelic effects are individually small, identification of these alleles is a daunting task. A mutation may not cause disease in a single molecular step, but instead interact with other mutations or cascade through intermediate molecular pathways eventually resulting in disease traits. Identifying the networks which underlie variation in disease traits and understanding how perturbations in these networks affect phenotypes may be a more efficient approach to identifying therapeutic targets. We propose to test this approach by building predictive genotype to phenotype models for complex behaviors in D. melanogaster and D.simulans (courtship and locomotion) assayed under standard conditions and during ethanol exposure (a model for environmental perturbations or disease states). By moving beyond descriptive studies in single populations to causative directional models which are predictive in novel conditions, shared and specific network structures will be identified. Networks which do not differ between sexes, species or environmental conditions are good candidates for comparison to other model systems or to humans. Highly conserved regulatory relationships are most likely critical for viability or fertility. Divergence in the regulation of individual genes or changes in network structure highlight the possible evolution of regulatory interactions among genes and identify candidate polymorphisms, genes and networks which may contribute to the evolution of novel traits, such as increased ethanol tolerance. Alcohol dependency affects approximately 18 million people in the USA, impacting health and quality of life. Ethanol has severe effects on human behavior, resulting in euphoria and sedation. In addition, intoxication can increase aggression, cause impotence, and decrease sexual inhibition and locomotor skills. The effects of alcohol and risk for alcoholism have a strong and complex genetic component, but are also mediated by the environment. While there are some large effect loci (e.g. ADH, ALDH and GABA receptor genes), much of the heritability remains unexplained. Drosophila have similar responses to ethanol, with courtship and locomotion affected by exposure. We will bring the power of Drosophila genetics to bear on this problem with an in depth examination of two species and their response to ethanol. We will directly test the hypothesis that conserved networks correspond to a higher degree of translatability of prediction of phenotypic outcomes across species. In addition, we will determine how robust these predictions are to gene by environment effects by exposing both populations to ethanol.